Developing nanoparticles to improve the specificity of anticancer agents towards tumor tissues and to better control drug delivery is a rising strategy in oncology. An increasing number of forms (e.g., conjugated nanoparticles, liposomes, immunoliposomes…) are now made available on the shelves and numerous other scaffolds (e.g., dendrimeres, nanospheres, squalenes…) are currently at various stages of development. However, the attrition rate when developing nanoparticles is particularly high and several promising forms showing excellent behavior and efficacy in preclinical studies failed to succeed in subsequent first-in-man studies or later in phase-II trials. The issue of pharmacokinetic variability is a major, yet largely underestimated issue with nanoparticles. A wide variety of causes (e.g; tumor type and disease staging, comorbidities, patient’s immune system) can explain this variability, which can in return impact negatively on pharmacodynamic endpoints such as lack of efficacy or severe toxicities. This review aims at covering the main causes for erratic pharmacokinetics observed with most nanoparticles. Should the main causes of such variability be identified, specific studies in non-clinical or clinical development stages could be undertaken using dedicated models (i.e., mechanistic or semi-mechanistic mathematical models such as PBPK approaches) to better describe nanoparticles pharmacokinetics and decipher PK/PD relationships. In addition, identifying relevant biomarkers or parameters likely to impact on nanoparticles pharmacokinetics would allow either modifying their characteristics to reduce the influence of the expected variability during development phases, or developing biomarker-based adaptive dosing strategies to maintain an optimal efficacy/toxicity balance. Overall, we call of developing comprehensive distribution studies and state-of-the-art modeling support to help better picture and anticipate nanoparticles pharmacokinetics.